Earlier studies have demonstrated that pure cultures of oral streptococci produce hydrogen peroxide but none has found any free peroxide in dental plaque or salivary sediment despite streptococci being major components of their mixed bacterial populations. The absence of peroxide in plaque and sediment could be due to the dominance of its destruction over its formation by bacterial constituents. To identify which of the oral bacteria might be involved in such a possibility, pure cultures of 27 different oral bacteria were surveyed (as well as dental plaque and sediment) for their peroxide-forming and peroxide-removing capabilities. Peroxide production was measured for each of the pure cultures by incubation with glucose at low and high substrate concentrations (2.8 and 28.0 mM) for 4 h and with the pH kept at 7.0 by a pH-stat. Removal of hydrogen peroxide was assessed in similar experiments where peroxide at 0, 29.4, 147.2 or 294.4 mM [0, 0.1, 0.5 and 1 % (w/v)] replaced the glucose. Hydrogen peroxide formation was seen with only three of the bacteria tested, Streptococcus sanguis I and II (sanguis and oralis), and Strep. mitior (mitis biotype I); levels of hydrogen peroxide between 2.2 and 9.8 mM were produced when these micro-organisms were grown aerobically and 1.1 and 3.9 mM when grown anaerobically. Earlier reports indicate that such levels were usually sufficient to inhibit the growth of many plaque bacteria. The amounts formed were similar at the two glucose levels tested, suggesting that maximum peroxide production is reached at low glucose concentration. None of the three peroxide-pro ducing organisms was able to utilize hydrogen peroxide but five of the other 24 tested, Neisseria sicca, Haemophilus segnis, H. parainfluenzae, Actinomyces viscosus and Staphylococcus epidermidis, could readily do so, as could the mixed bacteria in salivary sediment and dental plaque, both of which contain relatively high numbers of these peroxide-utilizing micro-organisms. The ability of the bacteria in plaque and sediment to degrade hydrogen peroxide was considerable and extremely rapid; peroxide removal was usually complete within the first 15 min of the incubation even when its initial level was as high as 294.4 mM. This almost overwhelming ability to remove peroxide was confirmed when peroxide-producing and -using cultures were mixed and when each of eight salivary sediments was incubated with glucose and with peroxide at concentrations up to 294.4 mM. In the glucose incubations, no hydrogen peroxide was observed, indicating dominance of microbial peroxide removers over hydrogen peroxide producers. In the peroxide incubations, the 294.4 mM concentration was the only one high enough to exceed the removal capacity of some of the sediments, a level that would only occur orally if provided in a therapeutic product. The optimum pH for the production of peroxide ranged between 4.0 and 7.0. The optimum pH for hydrogen peroxide removal ranged between 6.0 and 8.0. This suggests that for any peroxide accumulation to occur, an acidic rather than a neutral or alkaline pH might be more favourable. When glucose was replaced by galactose, a sugar readily available from salivary glycoprotein for oral microbial fermentation, the three peroxide-producing bacteria identified still produced peroxide but at slower rates than with glucose. Also, but unlike with glucose, more peroxide production was observed with 28.0 than with 2.8 mM galactose. These observations are consistent with fermentation of galactose being slower than that of glucose, and probably as a result the tendency to reach a less than maximum rate of peroxide formation with galactose than with glucose. When l(+) lactate, an intermediate or product formed in many oral bacterial fermentations, was supplied as substrate, only the streptococci identified in this study produced peroxide. The amounts and pattern of generation with all three micro-organisms were similar to those seen previously with glucose. Collectively, the findings were consistent with peroxide being produced when pyruvate, arising from either glucose or lactate, is converted aerobically to acetate and carbon dioxide and with this step being rate-limiting when glycolysis is rapid, as readily occurs with glucose but not with galactose.